RAM chip test

distropia · Sep 22, 2011

I'm trying to test 17 winbond 24258 chips (w24258s-70LE as used on mvs, also toshiba TC55257DFL-85L).

For that purpose, it is supposed I can use my eprom programmer (wellon vp280 with the last firmware/software updates) with a sop28 to dip28 adapter. I know is a small programmer, but in the specifications and software appears that chip as supported for testing.

The problem comes when I test it and every single chip says:

"Device Address:00004000H Device Data:FFH Buffer Address:00004000H Buffer Data:74H
>>Test fail!
>> 2,84Sec"

Always on the 4000H address (the buffer data changes from chip to chip). These chips are supposed to have 8000H, so it is weird it stops at the half capacity.

Any idea on how to test that chips or what could be wrong?

I was even thinking on soldering one by one in a mvs working board as "work ram" and make the workramtest, but the pcb could suffer soldering and desoldering 17 times on the same place.

Xian Xi · Sep 22, 2011

If it's stopping at half capacity and an address error that's probably the reason why it's at that capacity. Maybe the adapter isn't making good contact.

distropia · Sep 22, 2011

Xian Xi said:
If it's stopping at half capacity and an address error that's probably the reason why it's at that capacity. Maybe the adapter isn't making good contact.

I've received the adapter today, is brand new. The programmer makes a pin contact test before anything else and reports everything is OK (I simulated a bad pin and it was reported). I've programmed and verified 4 times an eprom to test the programmer socket (perfect and clean). The internal self test also marks ok. The chips are clean and correctly oriented, the same as the adapter... Don't know what else could I mention.

Xian Xi · Sep 22, 2011

Could be the firmware. With my burner if I use the newer version I get way more errors when testing or burning. Try using the version before the current one.

distropia · Sep 22, 2011

Xian Xi said:
Could be the firmware. With my burner if I use the newer version I get way more errors when testing or burning. Try using the version before the current one.

Thanks man. I've tried firmware v9 and 12 with the same result, the software updates the firmware but I can't choose the version I want. I've also tried with older soft version and the same occurs.

Is there any other way to test the ram?

EDIT: I just had an idea! I'll wire the adapter to a working MVS as work ram. That way the MVS will test it for me but without having to solder and desolder, just plug and test.

fremen · Sep 22, 2011

Hi,

Talking about the programmer firmwares, I bought recently a Xeltec 500 from buyicnow and I have had some problems, I downloaded the last version from Xeltec (USA) and it didn't work and the solution was to download the same version from Xeltec China and now the programmer works awesome.
The explanation from Xeltec to buyicnow was "a different policy from the company", this means that programmers from China only works with the "Chinese" version of the firmware nor with the "American" even if they are from the same manufacturer, so try the Chinese version (if it exists obviously) if your idea doesn't work.

Un saludo compadre 

channelmaniac · Sep 23, 2011

4000h is the bottom address of the top half of the chip.

If you look at it in binary with the addresses dividing by 2 every time you go down one you should look at the uppermost address line on the chip and make sure it's making good contact from the chip itself down to the pin that connects to the socket in the programmer.

Download the datasheet for the memory at http://www.datasheetarchive.com and see what the uppermost address line is.

Xian Xi · Sep 23, 2011

With adapters I find that after I clamp them I have to also push them straight down as well to get good contact.

Nightmare Tony · Sep 24, 2011

Possible JEDEC and non JEDEC pinout festivity since a high address and a chip select are swapped?

channelmaniac · Sep 24, 2011

Not sure Tony, I've never seen RAMs like that, but have seen plenty of 1M EPROMs and Mask ROMs with the screwy pinouts.

Nightmare Tony · Sep 25, 2011

The JEDEC and non JEDEC pinouts, think 27C1000 to 27C101 or 27C301. A16 and the output enable tend to get swapped...

distropia · Sep 26, 2011

http://www.datasheetarchive.com/dl/Datasheets-309/80633.pdf

Is that swapped?

Anyways my programmer supports the Winbond w24258/LL chip directly (I can choose it from the list). The only difference is the w24258S, but I guess it means SOP package.

I think the problem should be anywhere else.

Nightmare Tony · Sep 26, 2011

note sdure, look round for jedec pinout....(brain not running tonight)

Xian Xi · Oct 1, 2011

Am I doing something wrong?

I ordered a SOP 28 to DIP 28 adapter and I can't get the ram to sit in the clamp/socket or something.

This is the one I got:

http://www.ebay.com/itm/280594091224?ssPageName=STRK:MEWNX:IT&_trksid=p3984.m1439.l2649

Is that the correct one?

shadows · Oct 1, 2011

Are you pushing down on the socket? It's spring loaded and the pins will open up and clamp down once you let go of the socket.

distropia · Oct 1, 2011

Xian Xi said:
Am I doing something wrong?

I ordered a SOP 28 to DIP 28 adapter and I can't get the ram to sit in the clamp/socket or something.

This is the one I got:

http://www.ebay.com/itm/280594091224?ssPageName=STRK:MEWNX:IT&_trksid=p3984.m1439.l2649

Is that the correct one?

Man, exactly the same as mine. Even if you push down the socket up to the max, the chip does not fit inside. The adapter have a bad design. I already emailed the seller with pics showing the problem. I had to bend badly the chip legs to fit inside :very_ang:

Xian Xi · Oct 1, 2011

Ya, the ram chip itself is too wide to fit legs on both sides in, I also tried other brands and the same problem.

distropia · Oct 2, 2011

If it helps you, they answered me:

"Our programmer adapter is standard SO28 wide body (JEDEC MS-013 compliance which can be found from http://www.jedec.org/standards-documents/results/ms-013, registration needed).
You can also find the dimension from page http://www.fairchildsemi.com/packaging/surface_mount.html or link http://www.fairchildsemi.com/dwg/M2/M28B.pdf.
The standard descripe the wide Pin-Pin distance is 10.008~10.643mm.
And I tried the SOP28 IC on the adapter, it fits well.
The wide pin-pin distance of the W24258S is 0.453-0.477 inch (11.51-12.12mm).
So you can only try to bent the legs of the IC a little to fit into the adapter. Really very sorry for the troubles taken to you.
Later I will add much more description in my eBay page.
If the adapter really can't work for you, I would like to return back your money. It is also our fault, we don't describe very clearly in eBay page."

Xian Xi · Oct 2, 2011

Haha, same exact answer I got.

Ya the adapter can take only 10.008-10.643mm pin to pin but the static ram I have is windbond which is used on MVS and it is 11.81mm pin to pin.

RAM chip test

distropia

SouthTown StreetSweeper

Xian Xi

JammaNationX,

distropia

SouthTown StreetSweeper

Xian Xi

JammaNationX,

distropia

SouthTown StreetSweeper

fremen

Hardened Shock Trooper

channelmaniac

Mr Neo Fix-it

Xian Xi

JammaNationX,

Nightmare Tony

*Account control passed, on to Tony's family.Ex Ro

channelmaniac

Mr Neo Fix-it

Nightmare Tony

*Account control passed, on to Tony's family.Ex Ro

distropia

SouthTown StreetSweeper

Nightmare Tony

*Account control passed, on to Tony's family.Ex Ro

Xian Xi

JammaNationX,

shadows

Mature's Make-up Artist

distropia

SouthTown StreetSweeper

Xian Xi

JammaNationX,

distropia

SouthTown StreetSweeper

Xian Xi

JammaNationX,